Current issues of ACP Journal Club are published in Annals of Internal Medicine


Low income was an independent risk factor for premature death after controlling for health behaviors

ACP J Club. 1999 Jan-Feb;130:21. doi:10.7326/ACPJC-1999-130-1-021

Source Citation

Lantz PM, House JS, Lepkowski JM, et al. Socioeconomic factors, health behaviors, and mortality. Results from a nationally representative prospective study of US adults. JAMA. 1998 Jun 3;279:1703-8.



Do health behaviors explain increased mortality in poorer persons?


7.5-year cohort study (Americans' Changing Lives longitudinal survey).


United States.


3617 community-living adults who were ≥ 25 years of age (33% ≥ 65 y, 62% women, 32% black) and who lived in the United States (68% response rate). {All participants were included in the analysis.}*

Assessment of risk factors

Total years of education, income, and health behaviors (cigarette smoking, alcohol drinking, body weight, and level of physical activity). Age, sex, race, and residence were also assessed for potential confounding effects.

Main outcome measure

All-cause mortality data were obtained from the National Death Index and informants.

Main results

Poorer and less educated people were more likely to be smokers, sedentary, and overweight. After age, sex, and urban residence were adjusted for, the lowest income level (< $10 000/y) relative to the highest (≥ $30 000/y) showed an odds ratio of 3.13 (95% CI 1.97 to 4.95) for death in men and 3.82 (CI 1.86 to 7.85) in women. With further adjustment for smoking, alcohol, physical activity, and weight, the association with income remained statistically significant (Table). Other factors associated with mortality after adjustment for all other variables were age, sex, urban residence, being underweight, and low levels of physical activity (Table).


Adverse health behaviors were more common among people with low incomes. Low income remained a risk factor for premature death even after adjustment for the health behaviors of smoking, alcohol use, physical activity, and body weight.

Sources of funding: National Institute on Aging and Robert Wood Johnson Foundation.

For correspondence: Dr. P.M. Lantz, Department of Health Management and Policy, School of Public Health, University of Michigan, 109 Observatory, Ann Arbor, MI 48109-2029, USA. FAX 734-764-4338.

*Information on follow-up supplied by author.

Table. Risk factors for death in adults at 7.5-year mean follow-up

Risk factors Adjusted hazard rate ratio† (95% CI)
Age 35 to 44 y vs 25 to 34 y 2.66 (1.11 to 6.37)
Age 45 to 54 y vs 25 to 34 y 3.46 (1.20 to 9.95)
Age 55 to 64 y vs 25 to 34 y 9.30 (4.53 to 19.10)
Age 65 to 74 y vs 25 to 34 y 16.78 (8.17 to 34.47)
Age ≥ 75 vs 25 to 34 y 40.00 (19.10 to 83.93)
Women vs men 0.41 (0.30 to 0.54)
Urban vs rural residence 1.52 (1.10 to 2.10)
Annual income $10 000 to $29 999 vs ≥ $30 000 2.14 (1.38 to 3.25)
Annual income < $10 000 vs ≥ $30 000 2.77 (1.74 to 4.42)
Underweight vs normal body mass index 2.03 (1.32 to 3.12)
Physical activity, quintile 3 vs 5 (high) 1.60 (1.04 to 2.47)
Physical activity, quintile 2 vs 5 2.25 (1.41 to 3.58)
Physical activity, quintile 1 vs 5 2.91 (1.94 to 4.56)

†Hazard rate ratio adjusted for age, race, place of residence, sex, education, income, and health behaviors.


Lower socioeconomic status (SES) is one of the most powerful predictors of poor health and early death worldwide. High-risk health behaviors are more common in persons of lower SES. The excess mortality and morbidity associated with lower SES have been commonly attributed to differences in health behaviors.

Lantz and colleagues have shown that the increased prevalence of adverse health behaviors explained < 25% of the relation between lower SES and increased mortality. The study methods are strong. The investigators did not explore the factors associated with poverty that might explain the association with poor health, which include malnutrition, depression, chronic stress, poor housing, overcrowding, hostility, social isolation and lack of social support, and poor access to health care. Environmental factors, such as exposure to infectious diseases and pollutants, were also not evaluated. This does not take away from the important message of this study—that health-related behaviors explained very little of the excess mortality among persons with lower incomes and that other factors need to be considered.

The public health implications of the study findings are that interventions aimed at improving health behaviors among persons with low SES will have only a modest effect in reducing mortality and that efforts should be directed toward other factors that may be more successful in improving health outcomes. For practicing physicians, the findings of this study are probably not surprising. Most physicians see patients daily who confront enormous difficulties in their lives; it is hard to imagine that their health would not be adversely affected by these factors. This study also suggests that in addition to addressing lifestyle factors with patients, psychosocial factors, such as depression, loneliness, and social isolation, should be addressed when possible. Further, for physicians broadly interested in the health of their patients, the findings present a call to community, social, and political action.

Linda L. Humphrey, MD, MPH
Portland Veterans Affairs Medical CenterPortland, Oregon, USA